Re: Follow-up to determinist. error in lxml XML Schema validation

On 22/01/2012 18:20, Michael Rand wrote:
> David,
> This is a follow-up to the lxml schema import error question posted
> on StackOverflow (see URI below), to which you provided a response.
> As you suspected, eliminating the following line makes the schema
> import work without raising the same lxml exception:
> <xs:any namespace="##local" processContents="skip"/>
> Thanks for the response on this. Appreciate it.
> Mike Rand
> URI:

Thanks for the update.

For others following this list the lxml schema validator was giving
errors saying the MathML3 schema content model was non-deterministic,
I can't see why it should be, and the XSD validators that I tried did
not report any problems. The above change makes lxml happy (although
should not allow annotations in no namespace).

The content model of annotation-xml is slightly convoluted as it tries
to allow three things

* mathml annotations, which must be valid

* annotations in other namespaces, which need only be well formed (skip

* annotations in no namespaces, which need only be well formed (skip

logically the last two of these could be combined to a single "not
mathml namespace" choice but my reading of the XSD spec is that these
have to be specified separately as namespace="##local" and
namespace="##other" respectively. lxml appears to disagree.

If any other users have had error messages from XSD validation relating 
to "non deterministic" or "unique particle attribution" error 
conditions, could they let me know (on this list or directly)



The Numerical Algorithms Group Ltd is a company registered in England
and Wales with company number 1249803. The registered office is:
Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.

This e-mail has been scanned for all viruses by Star. The service is
powered by MessageLabs. 

Received on Monday, 23 January 2012 16:56:54 UTC